Optical character recognition (OCR) refers to the process of scanning or imaging printed, typewritten, handwritten, or other written text, and producing machine-recognizable text from the images. Due to imperfections in the documents, artifacts introduced during the imaging or scanning process, and shortcomings of OCR applications, errors often exist in the produced machine-recognizable text. Therefore, it would be desirable to efficiently identify and correct OCR errors and enhance the quality of text recognition.